Grand Theft Argument: the Problem with Paradigm Cases of Virtual Violence
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
The Philosophy of Computer Games Conference, Copenhagen 2018 Grand Theft Argument: the Problem with Paradigm Cases of Virtual Violence Rebecca Davnall 1. Does the player do wrong? Is virtual violence a natural kind? More importantly, is it an ethical kind? That is, are there ethical principles which generalise across all and only instances of something called ‘virtual violence’? This question is important because, if the answer is ‘yes’, then a reasonable foundation for virtual ethics will be the investigation of the relationship between virtual violence as an ethical kind and real violence as an ethical kind. On the other hand, if the answer is no, then the ethical picture will be much murkier – indeed, there may be no such thing as ‘virtual ethics’ at all. Most of the literature to date takes for granted a strong affirmative answer, holding that virtual violence is an ethical kind with a similar taxonomy to real violence. The root of this can be found in Morgan Luck’s definition of virtual murder: “A player commits an act of virtual murder in those cases where he directs his character to kill another in circumstances such that, were the game environment actual, the actions of his character would constitute actual murder.” (2008, p.31) Virtual murder then goes on to play an important role, as an ethical kind, in Luck’s discussion and much that has followed it. Tillson, for example, speaks quite openly of ‘immoral action types’ transcending an individual game’s frame: “I ask whether it is (as such) pro tanto wrong for a player to direct his character to perpetrate immoral action types within a videogame, and, more generally, whether it is (as such) pro tanto wrong for an individual to simulate perpetrating immoral action types.” (2018, p.1) In a somewhat overlooked paper, however, Seddon (2012) disagrees. He argues that the use of ‘virtual’ as a category term obscures more than it reveals, and that the ethical significance of a given virtual happening depends on the specific context of the simulation in which it occurs. Here, I will offer some concrete examples for the problem Seddon envisages. The view I am challenging works something like this: in the real world, actions can be individuated according to their physical descriptions (the throwing of a punch, firing a bullet into a body and so on), and virtual actions can be individuated by what physical events they resemble in the – paradigmatically audiovisual – output of the system they occur within. So if I do something while playing a game that makes one character appear to have shot another, I have brought it about that one character virtually shot another. This virtual shooting then draws whatever moral weight is attributed to it from the real-world moral import of really shooting someone. 1
Of course, really shooting someone may be thought to have different moral weight in different contexts. According to some ethical theories, shooting someone on a(n appropriately sanctioned) wartime battlefield is less wrong than shooting them in cold blood in peacetime. The view I am challenging can respect these kinds of variations; Luck, for example, specifically sets aside wartime killing because it would not legally count as murder in the real world (2008, p.32). This emphasises rather than diminishes the importance to this view of the connection between a virtual act and the real act it resembles. What I shall argue is that simulated acts do not draw their real moral weight (that is, the moral goodness or badness which attaches to the real human beings who cause them to be simulated) from the physical acts they represent. The distinction here may best be illustrated by a theatrical analogy: an actor playing a character who murders someone certainly simulates doing something wrong, and the wrongness of the act is part of the simulation, but the actor does not actually do something wrong in virtue of the wrongness simulated. The actor, as with the player of a violent game, may do something wrong in participating in the simulation, but they may also do no wrong in spite of simulating wrong. There are many ways that the actor could actually do wrong while simulating murder to which the content of the simulated act is irrelevant. For example, they could use the opportunity of a stabbing scene to actually hurt the actor playing the victim by failing to sufficiently pull their stroke. Or the theatre could be the target of a strike and the actor performing as a scab. Similarly, interacting with others through a game can be a vector for bullying and abuse; these acts are obviously wrong, but they are not my subject here. Instead, I shall argue that what is ethically important in representational acts quite generally is their symbolic power, and that this power depends too strongly on specific features of an individual simulation to be reducible to physical resemblance to real acts. What matters in a digital, interactive representation of murder is not that it resembles a real murder, but who is represented murdering whom, and what real social relationships are thus invoked. 2. Violent Symbols A useful comparison is the ethical assessment of speech-acts involving slurs. Efforts in the philosophy of language to account for the taboo around slur usage have struggled with the datum that groups targeted by a slur may reappropriate and make widespread use of the same word without taboo and yet leave the general taboo against the slur intact (see, for example, Hom 2008 and Kennedy 2003). One way to account for this is to say that the utterance of a slur by a member of that slur’s targeted group is, ethically, a different kind of act from the utterance of the same slur by someone outside the targeted group. How might this be the case? Well, a slur invokes a specific history of marginalisation. For example, most slurs targeting black people have their roots in the violent colonisation of African nations or the global slave trade that followed it. These were atrocities inflicted on black people by white people. When an anti-black slur is used by a white person (regardless of who hears it), the speech-act symbolically aligns with the history, and thus can be ethically linked to the ongoing harms which arise from that history. When a black person utters the same word, however, no symbolic link is established. This strategy can also be used to explain the fact that there are no slurs for groups which are not subject to histories of marginalisation. While there are sometimes words for these groups 2
which have some negative connotation – a common example is ‘limey’ to refer to the English – they have nothing like the cultural force of what are properly called slurs, because there is no history or present power dynamic for them to invoke. Much more would be needed to develop this into a mature account of the ethics of slurs. My point for now is merely to render plausible the claim that symbolic acts may have different symbolism, and thus different ethical weights, depending on who performs them. Another example of this might be ‘no-platforming’, the form of protest which aims to deny a speaker the authority and cultural capital of a particular venue or publication. No-platforming is clearly unethical when the targeted speaker belongs to a group which has historically been denied the opportunity to speak by the group doing the targeting (for example, when so called ‘men’s rights’ groups campaign to block feminist speakers reporting on misogynist abuse), since the no-platforming extends the harm of the history, both symbolically and in practice. When a marginalised group protest the speech of a privileged person calling for their further oppression, however, the protest runs counter to the history and thus may be entirely permissible. What this means for virtual violence is that ethical classification requires much more detailed attention to what’s going on in the simulation than the mere identification of its literal resemblance to real acts. The representation of a violent act within a game may be symbolic of something ethically problematic, but only in virtue of symbols embedded in the representation. Conversely, representations of actions that would, in real life, be entirely ethical may take on an objectionable symbolic character within a simulation, in virtue of how that simulation is situated in the real world1. In particular, it is unlikely that any single violent game can be taken as a paradigm case for all virtual violence. Even quite closely-related groups of games, such as war games, may vary widely in terms of what they symbolise. A game set during World War 2 plausibly carries a range of symbols (fascism, Nazism etc.) which a game set during World War 1 necessarily lacks; meanwhile the World War 1 game will directly involve the old empires of Europe whose forms radically shifted through the interbellum. The Modern Warfare strand of the Call of Duty franchise, because of its contemporary setting, will be at odds with the future-set Black Ops series. All war games symbolically invoke war and all its horrors, but different groups of people and patterns of behaviour are presented, in different ways and at different levels of abstraction, in each. Shooting a Nazi on the front lines in World War 2 is quite a different action to being in charge of a drone strike against a suspected terrorist hideout in a contemporary Middle Eastern village, for reasons that go far beyond the differences between the two physical acts. 3. Grand Theft Auto One game brand has come to serve, above all others, as paradigm case – and even rhetorical stand-in – for virtual violence: the Grand Theft Auto series. Since its inception in 1997, the series has been infamous for the wanton destruction and killing that players carry out during play. The games present an extremely violent image of criminal activity in America, both in 1 I don’t wish to develop this point in detail here, but a good example of this is the prevalence of white heroes and protagonists in Hollywood cinema; there is nothing wrong in the abstract with the hero of a film being white, but in the context of an industry where people of colour are pervasively denied fair opportunities and representation, every further white hero becomes a symbol of discriminatory exclusion. 3
the objectives they set for players and the affordances they make for non-goal-driven activity, the majority of which concern driving (especially into or over pedestrians) and shooting (especially of bystanders and police personnel). As well as becoming a byword for videogame violence, the Grand Theft Auto games are much-referenced in the academic literature. In addition to Luck (2009), references appear in Coeckelbergh (2007, p.219), Patridge (2010, p.311), Gooskens (2010, p.60), Nys (2010, p.84), Bartel (2011, p.13), Seddon (2012, p.3), Ali (2015, p.271 and p.273), Ostritch (2017, pp.123-4) and Tillson (2018, p.12). Not all of these papers discuss the games in depth, but all in one way or another take them as robust exemplars of virtual violence; as a case which any moral account of virtual violence ought to make neat work of. The question of whether it is wrong to play a Grand Theft Auto game turns, if what I have said above is right, on what symbolism it attaches to play actions and how those symbols are passed along or imputed to the player2. If the Grand Theft Auto games are to be paradigm cases of virtual violence, then, we should expect the ethically significant symbolism to be quite general. The franchise, however, stands at the intersection of at least three groups of symbolic links, which both encourage players to perform certain in-game acts and position those acts in relation to social structures that carry immense moral import. Some of these links are shared by other games; any game that indulges similarly in all three groups will be morally analogous to Grand Theft Auto. But I shall argue that the moral picture that emerges of the Grand Theft Auto franchise is complex and nuanced, and thus, by implication, that there are relatively few games that fit the same model. Firstly, the visual and narrative conventions of the Grand Theft Auto games are derived from pulp cinema and crime television. These are stories which tend to use violence as spectacle and to revel in the amorality of their characters, who are drawn especially from working-class backgrounds and are often the subject of racist stereotyping and caricature. Thanks to the global domination, for much of the 20th century, of Hollywood and the American media industry, players of Grand Theft Auto are likely to recognise the conventions of these genres and thus to see the games as expecting particular, violent, patterns of behaviour from the player character. Secondly, the games’ mechanics are set up to minimise the consequences of this behaviour. While police will respond to criminal acts carried out by the player, they are portrayed as laughably incompetent. Indeed, a whole style of play develops involving attracting and attempting to survive as much police attention as possible. If the in-game police do manage to arrest or even shoot the player character, the consequences are minimal; the character loses all his weapons and money, but is otherwise free to go about his business, reoffending at will. There’s no attempt on the part of the software to simulate criminal records and the heightened surveillance that marginalised arrestees are subjected to, nor the fracturing of community that an arrest or conviction can bring. The player is neither substantially punished for their in- game acts nor given any sense of what the real-life consequences of those acts would be. 2 In this I am setting aside the perennial question of whether playing violent games has long-lasting psychological effects which are ultimately harmful, though I accept that if they do, playing them will be wrong in proportion to the scale of the effect. For a summary of empirical work on the effects argument, see Anderson et al (2010) and Ferguson and Kilburn (2010). 4
Finally, there is the commercial and industrial context in which the games emerge. Over 1,000 people were involved in the creation of 2013’s Grand Theft Auto V (French, 2013), at an estimated total cost of $137.5million (Sinclair, 2013) to $256million (Villapaz, 2013); sales of the game grossed a billion dollars in three days (Goldfarb, 2013). Broadly speaking, the game affords players two kinds of activity: driving and violence. It is possible to drive responsibly, observing speed limits and traffic lights, but the majority of the development work that has gone into the driving has clearly been spent on higher-speed, more exhilarating, more dangerous vehicle use. A player buying the game is buying a violence machine; there is little point buying Grand Theft Auto V for other purposes, since there is (at least proportionately) little else to do in the game. The Grand Theft Auto games, then, both create an expectation in the player that the character through whom they interact with the virtual world should act violently – because the character is a criminal and because, well, what else is there for them to do? – and conspicuously minimise punishment for the character doing so. While it would be overstating matters to call this incitement, per se, it is nevertheless a form of encouragement. Particularly important is that, in presenting the consequences for violent acts as so thin, the games distance themselves from real-world intuitions about the moral weight of these acts. A particular representation of the world also emerges in the context of the games. While the cities in which most Grand Theft Auto games are set are fictional, they are close analogues of real-world cities (Grand Theft Auto IV’s Liberty City, for example, is clearly New York), and the specific cultures of those cities are invoked with varying degrees of subtlety. The characters in the game thus become, implicitly, representations of real people in real places. Since the majority of characters are working class, non-white and criminal, as well as often explosively violent, the games’ designers paint a picture of working class people of colour as dangerous criminals. The player then participates in bringing these representations ‘to life’. The result of all this is that, in engaging with a Grand Theft Auto game – loading up the software and interacting with it – the player is encouraged to perform a particular kind of representation of criminality and race. The presentation of working class people of colour as criminals has a long and sordid history of use to justify racist policies and actions, most recently highlighted by President Trump’s vilification of immigrants to the US as members of the gang MS-13. Grand Theft Auto, therefore, involves the strongest kind of symbolic linkage to real-world injustice. Who creates these representations matters, and while it is important to keep in mind that the majority of this labour is done by professional developers and designers (overwhelmingly white and male3), the player is the person located most directly in contact with these symbols as they are instantiated. The player chooses to play, and brings their own social position to that choice. There may be reappropriative ways for black people to engage with Grand Theft Auto games, analogous with the reappropriation of slurs, but as a white academic it is not my place to judge these. Any white person engaging with these games, though, must be subject to careful scrutiny for the ways in which they contribute to the symbolic reinforcement of racism and class. 3 Per IGDA (2017), 74% of people working in close relationship to the games industry are male and 64% are white, though this survey mainly surveyed people in western countries with majority-white populations. Perhaps more revealing in this context is that “Aboriginal or Indigenous people also represented 2% of the group, and Black/African American or African made up 1%” (pp.11-12). 5
4. Satire? It is worth digressing here a moment to address a defence, specifically of Grand Theft Auto, which has gained some currency in the literature. This is the argument, advanced by Ali (2015) and Ostritsch (2017), that these games are satirical, and thus the acts of violence carried out in them are insulated from the moral load they would otherwise bear. In our context, we can interpret this as the claim that the games’ satirical gloss of their events breaks (or at least sets at significant remove) the symbolic links discussed above. To describe the Grand Theft Auto games as satire is charitable. That they attempt humour is undeniable, but satire is a specific and quite narrow form of humour. I take it to have three distinct elements: an image of its target (i.e. a reproduction of what is to be satirised), an ironic distancing of the image from what it represents, which allows space for humour to occur, and a clear critique or argument about what is represented. It is not clear that the Grand Theft Auto games can manage all three simultaneously. Ali says the series “critiques and lampoons society’s violence and injustice” (2015 p.273); for Ostritsch, “the world of GTA V can easily be identified as a satirically exaggerated version of our world, the real one. But as satire GTA V does not endorse what it portrays, rather it ridicules it.” (2017 p.124). But what does this ridicule amount to? The games make no coherent argument about the structure of society; they suggest no alternate possibilities for organising life and sustaining community; indeed, by relying so heavily on racist cultural presuppositions, it is not clear that they represent ‘society’s violence and injustice’ or ‘the real [world]’ at all. Perhaps the idea is that, rather than critiquing the actual functions and behaviours of real social systems, the games target the mass-media images of crime and race that they present. But then the distance between image and target collapses. Grand Theft Auto games, on this model, are the very things they critique; indeed, given the scale of money and audience involved, they are one of the landmarks of this kind of imagery. There is little if any self- reflection in the games to countenance this latter version of the excuse. There is a further problem with the satire response, too, the problem of what Patridge (2010) calls ‘incorrigible social meanings’. In reference to an image of President Obama eating watermelon (the image of black people eating watermelon has strong associations to slavery in America), she writes: “[I]t would be very difficult for someone to use this kind of imagery in contemporary American culture in a way that avoids, or undermines its racist meaning. Such an interpretation would be, and it seems it ought to be, the first that occurs to those who are aware of this history. Even if the image’s author were to claim that the image was not meant to be interpreted as a racist insult but as a compliment, those who are adequately aware of the history of this kind of imagery and adequately sensitive to the current plight of African Americans will refuse to see the image as the author intends, even while recognising that the author’s intentions were otherwise. Thus, there are limits on what counts as a reasonable interpretation of such imagery.” (2010 p.308) 6
Perhaps the designers of the Grand Theft Auto games intend them as satire, but if they do, they do so with morally significant disregard for the history and context they are reinforcing. So too the player who plays the game without giving due consideration to their own personal social location relative to this kind of image. Patridge’s discussion of how incorrigible social meanings vary with social location is limited to how the meanings themselves might be contextual. For example, she argues, a representation of Obama eating watermelon that was created and consumed outside the context of American-dominated media would not have the same racist force, because the racial association of watermelon is specific to American culture. This may be true (although one has to go a long way to be out of the reach of American-dominated media), but it does not tell us anything about how different individuals may relate to these images in virtue of their social location. There is a clue in the above quote, however. Black Americans are much more likely to be ‘adequately aware of the history of this kind of imagery and adequately sensitive to the current plight of African Americans’, simply because it is almost impossible for them to avoid the latter and thus to see the former through it. Meanwhile the teaching of the history of slavery to white Americans (and British people) is an ideological battleground every bit as intense as the current-affairs controversy over whether modern majority-white nations are racist. Substantial effort is put in to modifying or obscuring the historical record to deny or minimise the harms of modern and historical racism. In linking the racist representations of Grand Theft Auto to current, real-world harms I do not mean to make a purely consequentialist case against the game. Indeed, Patridge is very clear that the moral fault in creating a representation that has an incorrigible social meaning is independent of its receipt by any audience. It may be that Grand Theft Auto’s staggering success has significantly promoted an image of American people of colour which has motivated acts of racist hatred and violence. Even if it has not, however, there is still a moral fault in creating (or instantiating through play) these representations, where one’s own relationship to what is represented is symbolically aligned with the representation. 5. Conclusion: things that are real I have argued that the moral fault in playing a violent game depends on whether parallels exist between the relationships represented as violent and the player’s relationship to the groups being represented. For example, Grand Theft Auto’s representation of working class people of colour as violent means that a white person playing the game participates in a representation that has historically been used by white people to justify racial discrimination against people of colour. The white player thus makes themselves by default complicit in this representation and its uses; only a substantive and careful effort to ensure that no such use is made can overturn this. A comparison case may be useful here. The indie game things that aren’t real (Vextro, 2017) includes a section which is, structurally, quite similar to freeform rampage in Grand Theft Auto. The player controls a character – in this case a minimally-developed, low-detail sprite – who moves around an area filled with identical non-player characters. The only interaction possible is for the player character to stab their neighbours; there is no reward for doing so, nor is it necessary to progress (the player can move to the next section at any time, though once there they cannot return). 7
This sequence is, as I understand it, a critique of how vacuous freeform violence can be in a game. I do not think, though, that playing the sequence is morally wrong, even if one actually does stab some of the non-player characters. This is because the character-tokens on screen are not sufficiently detailed to represent any particular real-life injustices; unlike the vast majority of real-world acts of violence, the killings in things that aren’t real really are motiveless, contextless acts. This account also shows where Tillson (2018) overstates matters in claiming that all video game violence is morally wrong because all violence involves a failure to respect sentience, and video game violence necessarily represents this lack of respect. The problem with this claim is that few if any human beings are in a social position such that their in-game actions can be symbolically aligned with a history of abuse of all sentient life (perhaps Donald Trump and a few other billionaire megacapitalists?) ‘Virtual violence’, ‘video game violence’ and their other loose cognates do not form clear ethical kinds. The ethical categories we should use to assess the creation and playing of violent games concern symbolism; how one’s own social position is symbolised by the engagement one has with a representation. Games GRAND THEFT AUTO IV. Rockstar, Xbox 360, 2008. GRAND THEFT AUTO V. Rockstar, Playstation 4, 2013. THINGS THAT AREN’T REAL. VEXTRO, PC, 2017. References Ali, R. 2015 A new solution to the gamer’s dilemma. Ethics and Information Technology 17: 267-274 Anderson, C.A. et al 2010 Violent Video Game Effects on Aggression, Empathy, and Prosocial Behavior in Eastern and Western Countries: A Meta-Analytic Review. Psychological Bulletin 136(2): 151-173 Bartel, C. 2012 Resolving the gamer’s dilemma. Ethics and Information Technology 14: 11- 16 Coeckelberg, M. 2007 Violent computer games, empathy and cosmopolitanism. Ethics and Information Technology 9: 219-231 Ferguson, C.J. and Kilburn, J. 2010 Much Ado About Nothing: The Misestimation and Overinterpretation of Violent Video Game Effects in Eastern and Western Nations: Comment on Anderson et al. (2010). Psychological Bulletin 136(2): 174-178 French, M. 2013 Inside Rockstar North – Part 2: The Studio. MCV [online] 4 October. Available at: (last accessed 23/07/2018) Goldfarb, A. 2013 GTA 5 sales hit $1 billion in three days. IGN [online] 20 September. Available at: (last accessed 23/7/2018) Gooskens, G. 2010 The Ethical Status of Virtual Actions. Ethical Perspectives 17(1): 59-78 Hom, C. 2008 The Semantics of Racial Epithets. The Journal of Philosophy 105(8): 416-440 8
Independent Game Developers Association (IGDA), 2018. Developer Satisfaction Survey [online] Available at: (last accessed 23/07/2018) Kennedy, R. 2003 Nigger: the strange career of a troublesome word. New York: Vintage Books Luck, M. 2009 The gamer’s dilemma: An analysis of the arguments for the moral distinction between virtual murder and virtual paedophilia. Ethics and Information Technology 11:31- 36 Nys, T. 2010 Virtual Ethics. Ethical Perspectives 17(1): 79-93 Ostritsch, S. 2017 The amoralist challenge to gaming and the gamer’s moral obligation. Ethics and Information Technology 19: 117-128 Patridge, S. 2010 The incorrigible social meaning of video game imagery. Ethics and Information Technology 13: 303-312 Seddon, R.F.J. 2012 Getting ‘virtual’ wrongs right. Ethics and Information Technology 15: 1-11 Sinclair, B. 2013 GTA V dev costs over $137 million, says analyst. gamesindustry.biz [online] 1 February. Available at: (last accessed 23/07/2018) Tillson, J. 2018 Is it distinctively wrong to simulate doing wrong? Ethics and Information Technology [online first] Villapaz, L. 2013 ‘GTA 5’ Costs $265 Million To Develop And Market, Making It The Most Expensive Video Game Ever Produced: Report. International Business Times [online] 8 September. Available at: (last accessed 23/07/2018) 9
You can also read